the 60Hz Harmonics Problem

1
the 60Hz Harmonics Problem

• Peter Cameron

2
Outline

• Old Stuff rehashing what almost everyone has
  seen at least once
• tell them what you are going to tell them, then
  tell them, then tell them what you told them
• New stuff
• another look at May APEX experiment data
• harmonics beyond the betatron line
• Plans for Run 7
• Common mode???
• Ramping supply at injection definitive?
• Repeat the APEX of last May
• DAQ

3
60Hz Harmonics

• direct excitation of betatron resonance by high
  harmonics of power supply frequencies
• mechanism is not yet understood
• large amplitude
• 40dB above 3D AFE noise floor during injection
  and store (1m)
• 70dB above 3D AFE noise floor during ramping
  (30m)
• three definitive measurements
• increase/decrease coincides with ramping supplies
• coupling moves harmonics into vertical
• main dipole 12 phase power supply tweak modifies
  spectrum
• Concerns
• makes BBQ/TF system refinement at RHIC difficult
• emittance growth???

4
the good news if 60Hz harmonic are not blowing
up emittance on the ramp (?), then BBQ certainly
wont
5
Effect of 60Hz Harmonics on BBQ
PLL locked on 60Hz harmonic
6
coupling variation shows excitation is in
horizontal plane
horizontal
vertical
7
12 phase power supply balancing

• At injection, tweak one phase of 12 phase power
  supply
• Causal relation was observed with pattern of 60Hz
  harmonics
• The tweak was large 150msec, or 3 degrees at
  60Hz
• Pattern changed, but total 60Hz harmonics power
  remained constant
• not likely to be eliminated by better balancing?
• problem is complex 12 free parameters, not
  reasonable to expect improvement from random
  tweaks
• fact that line amplitudes decreased suggests
  there is no other comparable source
• This phenomenon is not understood

8
NI DSA
755am 806am 807am
HP DSA
9
Correlations
correlation coefficients
determination coefficients

• Coefficient of Determination
• square of correlation coefficient
• more intuitively meaningful than correlation
  coefficient
• defined as the proportion of variation in common
  between two data sets
• with no noise, no machine drifts, perfect
  reproducibility of magnets would be 1 for
  755am and 807am data sets
• Random probability of observed determination
  coefficients is lt 1 part in 105

10
RHIC 6 phase 12 pulse
firing order A1, A2 B1, B2 A3, A4 B3, B4 A5,
A6 B5, B6 Phase between A and B bridges depends
on mains voltage (booster on/off, ???...) With
phase shift between A and B bridges, this
reproduces the pattern used in the simulation.
1 3 5
A
4 6 2
Df30 deg
1 3 5
B
4 6 2
11
3D AFE M Turn BPM Simulation
12
DC to 20KHz on the ramp
only 720Hz lines high-lighted
position of minimum determined by amount of SCR
overlap
note amplitude at betatron line amplitude at
720Hz according to million turn BPM
13
Outline

• Old Stuff rehashing what almost everyone has
  seen at least once
• tell them what you are going to tell them, then
  tell them, then tell them what you told them
• New stuff
• another look at May APEX experiment data
• harmonics beyond the betatron line
• Plans for Run 7
• Common mode???
• Ramping supply at injection definitive?
• Repeat APEX of last May
• DAQ

14
Baseband data from Carl Schultheiss
15

• APEX experiment baseline
• baseband data arbitrarily renormalized for
  display, renormalization the same for all
  graphs
• 60Hz harmonics at betatron line (h340) aligned
  with main dipole PS baseband spectrum
• spectra are quite different

• APEX experiment one phase tweaked
• power in betatron spectrum remains constant
• power in baseband spectrum goes up a lot (15dB
  on average)

16

• baseband data arbitrarily renormalized for
  display, renormalization the same for all
  graphs
• difference between spectra at baseband and
  betatron line with one phase tweaked
• baseline difference between spectra at baseband
  and betatron line

• difference between tweaked and baseline spectra
  at baseband
• power goes up a lot (15dB on average)
• difference between tweaked and baseline spectra
  at the betatron line

17
Conclusions May APEX experiment

• clear causal relation between main dipole PS
  phasing and 60Hz harmonics
• Simulation of 12 phase power supply and
  Comparison with Data
• imbalance between the two 6-phase bridges
  reproduces one observed spectral pattern
• overlap between phases reproduces observed null
  at 10KHz
• Comparison of baseband and betatron spectra
• they are quite different - what does this imply
  about the source? also see following slides
• power in baseband goes up a lot with single
  (large) tweak

18
Yellow horiz
span 5KHz 60Hz harmonics evident at 31KHz
Blue horiz
span 50KHz additional resonances in yellow, not
at the betatron line?
Yellow horiz
Yellow vert
19
spectral pattern a function of frequency overlaps
of 12 phase power supply?
proposed new p-p working pt
20dB enhancement at betatron line effective Q
of betatron resonance is 100?
Blue horiz span 39KHz, center 21KHz
20
Outline

• Old Stuff rehashing what almost everyone has
  seen at least once
• tell them what you are going to tell them, then
  tell them, then tell them what you told them
• New stuff
• another look at May APEX experiment data
• harmonics beyond the betatron line
• Plans for Run 7
• Common mode???
• Ramping supply at injection definitive?
• Repeat the APEX of last May
• DAQ

21
Plans for Run 7

• Carl additional filtering, gain 5dB
• Fritz Caspers - Common mode excitation??? will
  check (also capacitance to ground)
• Eduard - measure PS ripple AC couple, no
  magnetic core
• Ramping supply at injection definitive?
• George Ganetis/Kevin Brown trivial to switch to
  ramping supply while at injection
• look for changes in spectral pattern,
  amplitudes,
• Repeat last Mays APEX with more refinement,
  systematic attempt to tune out harmonics
• additional DAQ helpful presently just FFT

22
Bunch-by-bunch turn-by-turn system for Run 7
(Tom Russos talk tomorrow)
1m Stripline BH in BH out BV up BV down YH in YH
out YV up YV down
D hybrids
HP VME Mux
14 bit A/D ch1 ch2 clock/3 trigger
Control System GPMs, archiving,
SNAP
28MHz RF
local 1TB hard drive (archiving)
optional ext trigger (Artus,)
Windows PC running LabVIEW
23
Bunch-by-bunch turn-by-turn system for Run
7 modified to be sampling synchronous to
selectable 60Hz harmonic boxcar average the data
to improve S/N idea is to be able to observe
effect of phase tweaks in time domain
1m Stripline BH BV YH YV
3D AFE BH BV YH YV
14 bit A/D ch1 ch2 clock/n trigger
Control System GPMs, archiving,
SNAP
60Hz
n x h x PLL
local 1TB hard drive (archiving)
optional ext trigger
Windows PC running LabVIEW